In-cell touch oled frame synchronization multimodal sensing and intelligent inference system

Abstract

The application discloses a multi-modal sensing and intelligent reasoning system for frame synchronization of intra-cell touch OLED, and relates to the field of touch display driving chips; solves the problem that traditional chips need to be manually debugged screen by screen and cannot be real-time adapted to intra-cell touch OLED panel degradation; the application collects two types of analog signals of touch capacitance and pixel current in the display blanking area in time, performs cut-off modulation and demodulation on the touch signal to suppress display noise; after analog-digital conversion, statistical dispersion features and degradation deviation features are extracted according to preset physical blocks, and a fusion feature vector is generated in combination with the current display frame gray scale; a miniature multi-layer perception machine is used to identify the degradation mode, a non-linear compensation model is used to calculate a dynamic compensation voltage, and then a configuration instruction package is synthesized; the configuration instruction package is written into a shadow register in a reserved time slot, and takes effect in the next display frame. The application realizes frame-level real-time adaptive closed-loop control of the driving parameters and touch parameters of the intra-cell touch OLED panel.

Description

Technical Field

[0001] This invention belongs to the field of touch display driver chips, specifically a multimodal sensing and intelligent inference system for frame synchronization of touch OLED within a unit. Background Technology

[0002] In-cell touch technology refers to an integrated solution that combines touch sensor electrodes and display pixel driving electrodes using internal panel wiring. Compared to traditional overlay touch or external touch, in-cell touch structures can effectively reduce display module thickness, increase light transmittance, and lower manufacturing costs. In in-cell touch OLED panels, the same set of wiring performs display driving, touch sensing, and pixel status detection functions at different times. This multiplexing characteristic places higher demands on the timing control accuracy and signal processing capabilities of the driving chip.

[0003] Organic light-emitting diode (OLED) display technology is widely used in smartphones, tablets, and wearable devices due to its self-emissive, high-contrast, and fast-response characteristics. However, OLED pixels suffer from degradation during long-term operation, including threshold voltage drift and luminous efficiency decay. Different pixels exhibit varying degradation rates, leading to uneven brightness and color deviations. Traditional compensation methods rely on fixed parameter tables written at the factory or periodic offline calibration, making it difficult to track pixel degradation in real-time and perform frame-by-frame adaptive adjustments.

[0004] The following problems exist in the existing technology: In traditional solutions, touch signal acquisition is susceptible to interference from display driver noise and lacks effective analog domain noise reduction methods, resulting in a decrease in the touch detection signal-to-noise ratio. In existing technologies, the pixel driving current in intra-unit touch OLED panels deteriorates with operating time and temperature, and the degree of deterioration varies among different pixels. Fixed compensation parameters cannot track the deterioration state in real time. In existing solutions, the measured values ​​of touch capacitance are affected by changes in chip junction temperature, resulting in drift. The lack of a frame-level real-time temperature compensation mechanism leads to deviations in the touch detection threshold. Existing technologies lack a closed-loop control mechanism that completes the process from multimodal signal acquisition, feature extraction, inference decision-making to parameter updating within the blanking region of the same display frame, thus failing to achieve frame-by-frame adaptive adjustment. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a multimodal sensing and intelligent inference system for frame synchronization of in-cell touch OLED to solve the above-mentioned technical problems.

[0006] A first aspect of the present invention provides a frame-synchronized multimodal sensing and intelligent inference system for an in-cell touch OLED, comprising the following modules: The frame synchronization timing control module is used to respond to the vertical synchronization signal of the display timing controller, and sequentially trigger the touch sensing window and the pixel status sensing window within the display blanking area indicated by the signal, and the sum of the durations of the two windows is less than the preset ratio of the display blanking area; A multimodal sensing acquisition module is used to multiplex the traces of the touch OLED panel within the unit as touch receiving electrodes within the touch sensing window to acquire a first modal analog signal and perform chopping modulation and demodulation on the signal; and to multiplex the traces as detection electrodes within the pixel state sensing window to acquire a second modal analog signal. The on-chip feature extraction and fusion module is used to acquire the current junction temperature of the chip and perform temperature compensation on the first modal signal value obtained by quantizing the first modal analog signal according to a preset temperature coefficient. It also divides the first modal signal value and the second modal signal value obtained by quantizing the second modal analog signal into preset physical blocks, calculates the statistical deviation feature of the temperature-compensated first modal signal value relative to the historical baseline value in the non-volatile memory and the degradation deviation feature of the second modal signal value relative to the factory initial value in the non-volatile memory for each physical block, and fuses the statistical deviation feature and degradation deviation feature based on the grayscale value of the display frame corresponding to the vertical synchronization signal to generate a fused feature vector. The hybrid reasoning and compensation decision module incorporates a micro multilayer perceptron and a nonlinear compensation model. The micro multilayer perceptron obtains a degradation mode flag based on the fused feature vector. The nonlinear compensation model generates a dynamic compensation voltage value based on the average drive current of the second mode signal value, the temperature rise of the current junction temperature relative to a preset reference temperature, the degradation deviation characteristics, and the preset temperature acceleration factor. A configuration instruction package is generated based on the degradation mode flag and the dynamic compensation voltage value. The frame-level closed-loop update module is used to write the configuration instruction package into the display driver register and the touch threshold register within a reserved time slot other than the sum of the durations of the touch sensing window and the pixel state sensing window.

[0007] Preferably, the frame synchronization timing control module includes duration constraint judgment logic, which accumulates the configuration duration of the touch sensing window and the configuration duration of the pixel state sensing window, and compares the accumulated value with a preset ratio threshold of the display blanking area; when the accumulated value exceeds the preset ratio threshold, the configuration duration of the touch sensing window and the pixel state sensing window is truncated.

[0008] Preferably, the multimodal sensing acquisition module includes: An analog switch matrix has its input terminals connected to the traces of the touch OLED panel within the unit, and its output terminals connected to the touch sensing circuit within the touch sensing window to acquire the original first-mode analog signal, and connected to the pixel state detection circuit within the pixel state sensing window to acquire the second-mode analog signal; The touch sensing circuit includes a chopping modulation unit for modulating the original first-mode analog signal to a preset carrier frequency; and a synchronous demodulation unit for demodulating the modulated original first-mode analog signal using a reference signal that is in phase and frequency with the modulation end, and outputting the demodulated signal as the first-mode analog signal. The pixel state detection circuit includes a test signal generation unit for applying a test electrical signal to the trace within the pixel state sensing window to drive the pixel to conduct; and a transimpedance amplification unit for converting the current generated by the pixel conduction into a voltage signal to complete the acquisition of the second mode analog signal.

[0009] Preferably, in the on-chip feature extraction and fusion module, acquiring the current junction temperature of the chip and performing temperature compensation on the first modal signal value obtained by quantization of the first modal analog signal according to a preset temperature coefficient includes: Based on the current junction temperature of the chip, the preset reference temperature, the preset temperature coefficient, and the preset linear relationship between the first modal signal value and the first modal signal value, the temperature-compensated first modal signal value is calculated after the touch sensing window of the display frame corresponding to the vertical synchronization signal ends and before the statistical deviation feature is calculated; wherein the preset temperature coefficient is a coefficient value pre-calibrated for the capacitance temperature drift characteristics of the touch OLED panel traces in the unit.

[0010] Preferably, the on-chip feature extraction and fusion module further includes: The statistical deviation feature generation unit is used to perform a difference calculation between the standard deviation of the temperature-compensated first modal signal value in the display frame corresponding to the vertical synchronization signal and the standard deviation of the historical baseline collected under no-touch conditions during the chip initialization phase, to obtain the statistical deviation feature. The degradation deviation feature generation unit is used to calculate the root mean square value of the relative deviation of the second mode signal value with respect to the initial value collected under standard test conditions during the chip manufacturing stage, and to obtain the degradation deviation feature; The feature fusion unit is used to determine the weighting coefficient based on the grayscale value of each physical block corresponding to the display frame corresponding to the vertical synchronization signal, and to fuse the difference of the standard deviation and the root mean square value of the relative deviation with the weighting coefficient to obtain the fused feature vector.

[0011] Preferably, the hybrid reasoning and compensation decision-making module includes: The miniature multilayer perceptron consists of a fully connected input layer, hidden layer, and output layer. Its weight parameters are pre-programmed in a one-time programmable memory. The miniature multilayer perceptron is configured in a hardware accelerator and performs forward inference on the fused feature vector within the reserved processing time slot of the display frame corresponding to the vertical synchronization signal, and outputs a degradation mode flag. A nonlinear compensation model generates a dynamic compensation voltage value based on the average drive current of the second mode signal value, the temperature rise of the current junction temperature relative to a preset reference temperature, the degradation deviation characteristics, and the preset temperature acceleration factor. The instruction synthesis unit selects the corresponding gamma correction curve index value from the preset gamma correction curve lookup table according to the degradation mode flag, and combines the gamma correction curve index value with the dynamic compensation voltage value to form a configuration instruction package.

[0012] Preferably, in the hybrid reasoning and compensation decision module, the nonlinear compensation model generates dynamic compensation voltage values. The formula is: in, The average drive current of the second mode signal value, The factory-calibrated reference current value, The temperature rise of the current junction temperature relative to the preset reference temperature. Here, N is the preset temperature acceleration factor, and N is the number of pixels contained within the preset physical block. The current driving current value of the i-th pixel in the second modal signal value. The initial current value of the i-th pixel is the factory current value. This is the scaling factor constant.

[0013] Preferably, the frame-level closed-loop update module includes: The bus interface control unit is used to write the configuration instruction packet into the shadow registers of the display driver register group and the touch threshold register group via the internal chip bus within the reserved time slot; The register update control unit is used to load the contents of the shadow register into the working registers of the display driver register group and the touch threshold register group in response to the register update pulse signal generated by the display timing controller when the effective edge of the next vertical synchronization signal after the vertical synchronization signal arrives.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This invention uses a multimodal sensing acquisition module to perform chopping modulation and demodulation on the first-mode analog signal, thereby suppressing display drive common-mode noise in the analog domain and improving the touch detection signal-to-noise ratio. This invention uses a miniature multilayer perceptron in the hybrid reasoning and compensation decision module to output a degradation mode flag, which classifies pixel degradation states in real time. It also generates dynamic compensation voltage values ​​through a nonlinear compensation model, thereby achieving accurate voltage compensation for different degrees of degradation. This invention uses an on-chip feature extraction and fusion module to perform linear temperature compensation based on a preset temperature coefficient on the first mode signal value, thereby eliminating capacitance measurement drift caused by chip junction temperature changes and improving the stability of touch detection threshold. This invention uses a frame synchronization timing control module to trigger the touch sensing window and pixel state sensing window in a time-division manner within the display blanking zone, and sets duration constraint judgment logic to ensure that sensing and inference operations do not encroach on display driving time. Combined with a frame-level closed-loop update module, the configuration instruction packet is written into the shadow register within the reserved time slot and takes effect in the next display frame, realizing real-time closed-loop control from sensing to parameter update within a single frame cycle. There is no need to export the original data, reducing bandwidth usage and processing latency. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the module flow of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0017] Please see Figure 1 This invention is a frame-synchronized multimodal sensing and intelligent inference system for in-cell touch OLEDs, comprising the following modules: The frame synchronization timing control module is used to respond to the vertical synchronization signal of the display timing controller, and sequentially trigger the touch sensing window and the pixel status sensing window within the display blanking area indicated by the signal, and the sum of the durations of the two windows is less than the preset ratio of the display blanking area; A multimodal sensing acquisition module is used to multiplex the traces of the touch OLED panel within the unit as touch receiving electrodes within the touch sensing window to acquire a first modal analog signal and perform chopping modulation and demodulation on the signal; and to multiplex the traces as detection electrodes within the pixel state sensing window to acquire a second modal analog signal. The on-chip feature extraction and fusion module is used to acquire the current junction temperature of the chip and perform temperature compensation on the first modal signal value obtained by quantizing the first modal analog signal according to a preset temperature coefficient. It also divides the first modal signal value and the second modal signal value obtained by quantizing the second modal analog signal into preset physical blocks, calculates the statistical deviation feature of the temperature-compensated first modal signal value relative to the historical baseline value in the non-volatile memory and the degradation deviation feature of the second modal signal value relative to the factory initial value in the non-volatile memory for each physical block, and fuses the statistical deviation feature and degradation deviation feature based on the grayscale value of the display frame corresponding to the vertical synchronization signal to generate a fused feature vector. The hybrid reasoning and compensation decision module incorporates a micro multilayer perceptron and a nonlinear compensation model. The micro multilayer perceptron obtains a degradation mode flag based on the fused feature vector. The nonlinear compensation model generates a dynamic compensation voltage value based on the average drive current of the second mode signal value, the temperature rise of the current junction temperature relative to a preset reference temperature, the degradation deviation characteristics, and the preset temperature acceleration factor. A configuration instruction package is generated based on the degradation mode flag and the dynamic compensation voltage value. The frame-level closed-loop update module is used to write the configuration instruction package into the display driver register and the touch threshold register within a reserved time slot other than the sum of the durations of the touch sensing window and the pixel state sensing window.

[0018] Specifically, the frame synchronization timing control module receives the vertical synchronization signal from the display timing controller. Within the display blanking area indicated by the vertical synchronization signal, it sequentially triggers the touch sensing window and the pixel state sensing window, and the sum of the durations of the two windows is less than a preset proportion of the display blanking area to reserve processing time slots for subsequent inference calculations and register updates. Within the touch sensing window, the multimodal sensing acquisition module multiplexes the traces of the touch OLED panel within the unit as touch receiving electrodes through an analog switch matrix, acquires the first modal analog signal, and performs chopping modulation and demodulation on the first modal analog signal to suppress display driving noise. Within the pixel state sensing window, the multimodal sensing acquisition module multiplexes the traces as detection electrodes through an analog switch matrix, applies a test signal to turn on the pixel, and converts the generated current into a voltage signal to acquire the second modal analog signal. The on-chip feature extraction and fusion module quantizes the first modal analog signal into a first modal signal value and the second modal analog signal into a second modal signal value through an analog-to-digital converter. This module reads the current junction temperature measured by the chip's built-in temperature sensor and performs temperature compensation on the first mode signal value according to a preset temperature coefficient to eliminate the influence of temperature drift on capacitance measurement. The module divides the first and second mode signal values ​​into preset physical blocks. For each physical block, it calculates the statistical deviation characteristics of the temperature-compensated first mode signal value relative to the historical baseline value in non-volatile memory, and the degradation deviation characteristics of the second mode signal value relative to the factory initial value in non-volatile memory. It then fuses the statistical deviation characteristics and degradation deviation characteristics based on the grayscale values ​​of the current display frame corresponding to each physical block to generate a fused feature vector. The hybrid inference and compensation decision module incorporates a micro multilayer perceptron and a nonlinear compensation model. The micro multilayer perceptron receives the fused feature vector and performs forward inference, outputting a flag representing the panel degradation mode. The nonlinear compensation model generates a dynamic compensation voltage value based on the average drive current of the second mode signal value, the temperature rise of the current junction temperature relative to the preset reference temperature, the degradation deviation characteristics, and a preset temperature acceleration factor. This module generates a configuration instruction package based on the degradation mode flag and the dynamic compensation voltage value. Within a reserved time slot beyond the sum of the durations of the touch sensing window and the pixel state sensing window, the frame-level closed-loop update module writes configuration instruction packets into the shadow registers of the display driver register group and the touch threshold register group via the chip's internal bus. When the effective edge of the next vertical synchronization signal arrives, in response to the register update pulse signal generated by the display timing controller, the contents of the shadow register are loaded into the working register. The updated parameters automatically take effect in the next display frame, completing frame-level closed-loop control from signal acquisition, feature fusion, intelligent inference to adaptive adjustment of panel driver parameters.

[0019] In one embodiment of the present invention, the frame synchronization timing control module includes a duration constraint judgment logic, which accumulates the configuration duration of the touch sensing window and the configuration duration of the pixel state sensing window, and compares the accumulated value with a preset ratio threshold of the display blanking area; when the accumulated value exceeds the preset ratio threshold, the configuration duration of the touch sensing window and the pixel state sensing window is truncated.

[0020] Specifically, the frame synchronization timing control module receives the vertical synchronization signal from the display timing controller through the chip's internal interconnect interface. The vertical synchronization signal is a periodic digital pulse signal, and the time interval between two adjacent rising edges is defined as a complete display frame cycle; the display frame corresponding to the current vertical synchronization signal cycle is called the current display frame. The rising edge of the vertical synchronization signal triggers the blanking area positioning logic inside the frame synchronization timing control module; the blanking area positioning logic determines the start and end times of the display blanking interval of the current display frame based on the line scan counter status provided by the display timing controller; the display blanking interval is the time period from after the touch OLED panel in the unit completes writing the last row of pixel data for the current display frame to before the first row of pixel data for the next frame begins to be written; the touch OLED panel in the unit is an integrated display panel that shares internal wiring between the touch sensor electrodes and the OLED pixel driving electrodes, and its typical feature is that the same wiring can be used for display driving, touch sensing, and pixel status detection at different times.

[0021] The configuration register set inside the frame synchronization timing control module is written to the touch sensing window duration by the system firmware during the initialization phase via the chip system bus. Duration of pixel state sensing window . The value must be greater than the sum of the touch analog front-end settling time and the full array scan time of the analog-to-digital converter; in a typical implementation using a 64-channel successive approximation analog-to-digital converter with a sampling rate of 1 MHz, completing one full array capacitive scan takes approximately 0.8 to 1.2 milliseconds, requiring a margin of safety. The value ranges from 1.2 milliseconds to 1.8 milliseconds. The value must meet the settling time requirement of pixel drive current measurement. Under a 64×64 physical block partitioning scheme, if a parallel measurement architecture is adopted, it takes approximately 0.5 to 0.8 milliseconds to complete the sensing of all physical blocks. Therefore... The value ranges from 0.8 milliseconds to 1.2 milliseconds.

[0022] The frame synchronization timing control module internally includes a first timer, a second timer, and corresponding first and second comparators. The input of the first comparator is connected to the current count value of the first timer and... Connection, when the first timer count value is less than At this time, the first comparator outputs a valid level as the first enable signal. The input of the second comparator is then compared with the current count value of the second timer. Connection, when the second timer count value is less than When the second comparator outputs a valid level as a second enable signal, at the start of the display blanking interval of the current display frame, the first timer and the second timer are started sequentially. When the first timer is started, the first comparator reaches the first timer count value. The first enable signal that has been continuously outputting a valid level; when the first timer count value equals When the first comparator sets the first enable signal to an invalid level, it simultaneously starts the second timer. The second comparator then activates the second timer when the count value reaches [a certain threshold]. The second enable signal, which previously continuously output a valid level, will be active when the second timer count value equals... At that time, the second comparator sets the second enable signal to an invalid level.

[0023] The duration constraint judgment logic consists of an adder, a comparator, and a threshold register. The adder obtains the total window duration. The threshold register stores the preset ratio threshold for displaying the blanking area. The preset ratio threshold is defined as the maximum allowable proportion of the sum of the durations of the trigger touch sensing window and the pixel state sensing window to the total duration of the display blanking area. In a preferred embodiment, the preset ratio threshold is set to 80%, based on the fact that the time required for the micro multilayer perceptron in the hybrid inference and compensation decision module to complete one forward inference is less than 100 microseconds, the calculation of the nonlinear compensation model can be completed in parallel hardware, the threshold register write operation is completed in about 50 microseconds through the internal high-speed bus, and the 20% display blanking area duration is reserved as a processing time slot to provide sufficient timing margin.

[0024] The comparator compares the total window duration with a dynamic threshold. The actual total duration of the hidden area is displayed. The blanking area positioning logic calculates the value in real time based on the row scan counter status, and the dynamic threshold automatically adapts to changes in the blanking area length at different refresh rates or resolutions. Greater than Then the duration constraint judgment logic triggers a truncation operation; the truncation operation specifically involves... and The touch sensing window is scaled down proportionally to the original size, so that the sum of the duration of the scaled-down touch sensing window and the duration of the pixel state sensing window equals the dynamic threshold; the duration of the scaled-down touch sensing window... Duration of the reduced pixel state sensing window .like Less than or equal to If the duration constraint judgment logic does not intervene, it will be directly adopted. and The duration of the window for actual execution.

[0025] In one embodiment of the present invention, the multimodal sensing acquisition module includes: An analog switch matrix has its input terminals connected to the traces of the touch OLED panel within the unit, and its output terminals connected to the touch sensing circuit within the touch sensing window to acquire the original first-mode analog signal, and connected to the pixel state detection circuit within the pixel state sensing window to acquire the second-mode analog signal; The touch sensing circuit includes a chopping modulation unit for modulating the original first-mode analog signal to a preset carrier frequency; and a synchronous demodulation unit for demodulating the modulated original first-mode analog signal using a reference signal that is in phase and frequency with the modulation end, and outputting the demodulated signal as the first-mode analog signal. The pixel state detection circuit includes a test signal generation unit for applying a test electrical signal to the trace within the pixel state sensing window to drive the pixel to conduct; and a transimpedance amplification unit for converting the current generated by the pixel conduction into a voltage signal to complete the acquisition of the second mode analog signal.

[0026] Specifically, the input terminals of the analog switch matrix are connected one by one to the traces of the touch OLED panel within the unit, and the control terminals receive the first enable signal and the second enable signal respectively. When the first enable signal is at an active level, the analog switch matrix connects the traces of the touch OLED panel within the unit to the touch sensing circuit, at which time the trace is multiplexed as a touch receiving electrode; when a conductive object approaches or contacts the surface of the touch OLED panel within the unit, a coupling capacitance is formed between the touch receiving electrode and the conductive object, causing a change in the amount of charge on the touch receiving electrode, thereby generating a voltage signal with a corresponding amplitude change, which is the original first mode analog signal.

[0027] The chopping modulation unit in the touch sensing circuit contains a local oscillator that generates a modulation signal with a carrier frequency of 100 kHz to 500 kHz, selected to avoid harmonic interference from the display drive line scan frequency. It multiplies the original first-mode analog signal with the modulation signal, shifting the spectrum of the original first-mode analog signal to near the carrier frequency. The synchronous demodulation unit generates a reference signal with the same frequency and phase as the modulation signal, performs multiplication demodulation on the modulated original first-mode analog signal, shifting its spectrum back to baseband. The demodulated original first-mode analog signal is then passed through a low-pass filter with a cutoff frequency of 50 kHz to remove high-frequency components before being output as the first-mode analog signal.

[0028] Each pixel in the touch OLED panel within the unit is composed of an organic light-emitting diode (OLED). The display driving circuit is a circuit module that provides driving current to each pixel during the normal display phase, and is connected to each pixel through the traces of the touch OLED panel within the unit. During the touch sensing window and pixel state sensing window, an analog switch matrix temporarily disconnects the traces from the display driving circuit and switches them to the touch sensing circuit or the pixel state detection circuit, respectively. After the window ends, the traces are reconnected to the display driving circuit, and normal display driving continues.

[0029] When the second enable signal is active, the analog switch matrix connects the traces of the touch OLED panel within the cell to the pixel state detection circuit, at which point the traces are multiplexed as detection electrodes. The test signal generation unit in the pixel state detection circuit applies a test voltage signal to the selected traces of the touch OLED panel within the cell within the pixel state sensing window. The test voltage amplitude ranges from 3 volts to 5 volts, based on the typical anode-cathode voltage drop of a pixel composed of organic light-emitting diodes in the touch OLED panel within the cell when it is in the on state. This test voltage drives the pixel composed of organic light-emitting diodes in the touch OLED panel connected to the traces to enter the on state, generating a driving current. In the transimpedance amplifier unit, the feedback resistor is connected between the output terminal and the inverting input terminal of the operational amplifier. According to the virtual short and virtual open characteristics of the operational amplifier, the output voltage signal is the product of the drive current and the feedback resistor. This voltage signal is the second mode analog signal. The value of the feedback resistor ranges from 100 kΩ to 1 MΩ. This value range is based on the matching relationship between the typical pixel drive current range of 1 μA to 10 μA and the desired output voltage range of 0.5 V to 5 V.

[0030] In one embodiment of the present invention, the on-chip feature extraction and fusion module acquires the current junction temperature of the chip and performs temperature compensation on the first modal signal value obtained by quantization of the first modal analog signal according to a preset temperature coefficient, including: Based on the current junction temperature of the chip, the preset reference temperature, the preset temperature coefficient, and the preset linear relationship between the first modal signal value and the first modal signal value, the temperature-compensated first modal signal value is calculated after the touch sensing window of the display frame corresponding to the vertical synchronization signal ends and before the statistical deviation feature is calculated; wherein the preset temperature coefficient is a coefficient value pre-calibrated for the capacitance temperature drift characteristics of the touch OLED panel traces in the unit.

[0031] Specifically, the on-chip feature extraction and fusion module includes a 10-bit successive approximation analog-to-digital converter (ADC) with multiple input channels and a sampling rate of 1 trillion times per second. After the touch sensing window ends, the first input channel of the ADC performs sample-and-hold and quantization conversion on the first modal analog signal, converting the analog voltage value into a 10-bit digital value, which is the first modal signal value before temperature compensation. After the pixel state sensing window ends, the second input channel of the analog-to-digital converter performs sample-and-hold and quantization conversion on the second mode analog signal, converting the analog voltage value into a 10-bit digital value, which is the second mode signal value.

[0032] The temperature compensation step is performed after the touch sensing window of the display frame corresponding to each vertical synchronization signal ends and before the statistical deviation feature is calculated. The current junction temperature value is read from the chip's built-in temperature sensor. And read the preset reference temperature from non-volatile memory. and preset temperature coefficient . The value is 25 degrees Celsius, which is the ambient temperature used when the chip was calibrated at the factory. The capacitance temperature drift characteristics of the touch OLED panel traces within the unit are characterized. The physical source of this is the combined effect of the temperature coefficient of resistivity of the trace metal material and the temperature coefficient of dielectric constant of the interlayer dielectric material, which leads to an increase in the parasitic capacitance of the touch electrode as the temperature rises. The value of was obtained by measuring the capacitance of the touch OLED panel sample within the unit at different temperatures and then performing linear fitting; its range is as follows: Typical value The temperature-compensated first mode signal value is calculated based on a preset linear relationship. .For example, It is 25 degrees Celsius. for , It is 55 degrees Celsius. If the value is 1200, then the temperature rise is 30 degrees Celsius. The value is approximately 1189.3, eliminating the drift in capacitance measurements caused by temperature increases.

[0033] In one specific embodiment, during the wafer-level testing before chip delivery or the final testing stage after packaging, the ambient temperature of the chip is controlled to be... (For example, at 25°C), under non-touch conditions, the original first modal signal values ​​of each touch channel are read by the multimodal sensing acquisition module and recorded as the first reference value. Subsequently, the ambient temperature was raised to the second calibration temperature. (e.g., 55°C) After the chip junction temperature stabilizes, read the original first mode signal values ​​of each touch channel again and record them as the second calibration values. Finally, based on the linear relationship model... The preset temperature coefficient was obtained by fitting the measurement data from multiple channels using the least squares method. .

[0034] In one embodiment of the present invention, the on-chip feature extraction and fusion module further includes: The statistical deviation feature generation unit is used to perform a difference calculation between the standard deviation of the temperature-compensated first modal signal value in the display frame corresponding to the vertical synchronization signal and the standard deviation of the historical baseline collected under no-touch conditions during the chip initialization phase, to obtain the statistical deviation feature. The degradation deviation feature generation unit is used to calculate the root mean square value of the relative deviation of the second mode signal value with respect to the initial value collected under standard test conditions during the chip manufacturing stage, and to obtain the degradation deviation feature; The feature fusion unit is used to determine the weighting coefficient based on the grayscale value of each physical block corresponding to the display frame corresponding to the vertical synchronization signal, and to fuse the difference of the standard deviation and the root mean square value of the relative deviation with the weighting coefficient to obtain the fused feature vector.

[0035] Specifically, the on-chip feature extraction and fusion module logically groups the touch electrode array and pixel array of the touch OLED panel within the unit according to a preset physical block division scheme. Each electrode intersection in the touch electrode array corresponds to a sensing unit, and each sensing unit outputs a first modal signal value. In a typical embodiment, each physical block corresponds to a 16×16 pixel area on the display panel, that is, each physical block contains 256 first modal signal values ​​and 256 second modal signal values.

[0036] The statistical deviation feature generation unit is based on the first mode signal values ​​after temperature compensation within each physical block, denoted as... Where k ranges from 1 to M, and M is the number of sensing units within the physical block. In a 16×16 partitioning scheme, M=256. Calculate the standard deviation of the display frame corresponding to the current vertical synchronization signal for this physical block. ,in This is the average of all first-mode signal values. Then, the historical baseline standard deviation corresponding to this physical block is read from non-volatile memory. The historical baseline standard deviation was collected and stored during the chip initialization phase under non-touch and preset temperature conditions. The preset temperature condition was 25 degrees Celsius, consistent with the preset reference temperature, to ensure the baseline is comparable to the current measurement value. The historical baseline standard deviation reflects the touch noise floor level of this physical block under non-touch and standard temperature conditions. The statistical deviation characteristics of this physical block are thus obtained. , A positive value indicates that the noise level of the current frame is higher than the baseline. A negative value indicates that the noise level is below the baseline. The statistical deviation characteristics of all physical blocks constitute the statistical deviation feature vector.

[0037] The degradation deviation characteristic generation unit reads the initial factory current value corresponding to the pixels composed of organic light-emitting diodes in the touch OLED panel of each unit within the physical block from the non-volatile memory, and records it as... The initial current value is collected and stored during the chip manufacturing process under standard test conditions. These conditions include a typical operating voltage, an ambient temperature of 25 degrees Celsius, and a pure white test screen. Then, based on all second-mode signal values ​​within this physical block, it is denoted as... Generate the degradation deviation characteristics of this physical block ,in, This represents the k-th second-mode signal value within the physical block. This represents the k-th initial current value within the physical block. The degradation deviation characteristic reflects the spatial non-uniformity of the pixel driving current (composed of organic light-emitting diodes) in the touch OLED panel within the unit of this physical block relative to the factory state. The larger the D value, the more significant the inconsistency in pixel degradation within the physical block. The degradation deviation characteristics of all physical blocks constitute the degradation deviation feature vector.

[0038] The feature fusion unit obtains the grayscale values ​​of each physical block corresponding to the display frame of the current vertical synchronization signal from the display timing controller. The grayscale value represents the average brightness level of the display area covered by the physical block, and its value ranges from 0 to 255. A weighting coefficient w is determined based on the grayscale value of each physical block. The weighting coefficient w and the grayscale value G have a monotonically increasing relationship. As an optional implementation, w and G have a linear mapping relationship. As another alternative implementation, considering the gamma characteristics of the display panel, a non-linear mapping relationship can also be used, for example... Alternatively, a custom mapping can be implemented using a lookup table. Higher grayscale values ​​result in larger weighting coefficients; these coefficients adjust the contribution of statistical deviation features to the fusion features. The physical basis for this is that the display drive current amplitude increases with increasing grayscale value. Under high grayscale conditions, display noise has a stronger coupling interference on the touch electrodes, thus requiring higher weights for statistical deviation features to compensate for the impact of display noise on touch detection. The statistical deviation features and degradation deviation features of each physical block are weighted and concatenated to generate the fusion feature component of that physical block. A typical concatenation method involves arranging the weighted statistical deviation features and degradation deviation features sequentially to form a two-dimensional feature component. The fusion feature components of all physical blocks are concatenated in a preset order to generate a one-dimensional fusion feature vector. In a typical embodiment with 64 physical blocks (corresponding to an 8×8 physical block division), each physical block generates a two-dimensional feature component, and the total dimension of the fusion feature vector is 128.

[0039] In one embodiment of the present invention, the hybrid reasoning and compensation decision module includes: The miniature multilayer perceptron consists of a fully connected input layer, hidden layer, and output layer. Its weight parameters are pre-programmed in a one-time programmable memory. The miniature multilayer perceptron is configured in a hardware accelerator and performs forward inference on the fused feature vector within the reserved processing time slot of the display frame corresponding to the vertical synchronization signal, and outputs a degradation mode flag. A nonlinear compensation model generates a dynamic compensation voltage value based on the average drive current of the second mode signal value, the temperature rise of the current junction temperature relative to a preset reference temperature, the degradation deviation characteristics, and the preset temperature acceleration factor. The instruction synthesis unit selects the corresponding gamma correction curve index value from the preset gamma correction curve lookup table according to the degradation mode flag, and combines the gamma correction curve index value with the dynamic compensation voltage value to form a configuration instruction package.

[0040] Specifically, the miniature multilayer perceptron consists of a fully connected input layer, hidden layers, and an output layer. The number of neurons in the input layer is equal to the dimension of the fused feature vector. There are two hidden layers: the first hidden layer contains 64 neurons, and the second hidden layer contains 32 neurons. The output layer contains 16 neurons, corresponding to 16 preset panel degradation mode categories. The connection weight parameters between each layer are pre-trained and stored in a one-time programmable memory during chip manufacturing, and are automatically loaded into the weight register group of the miniature multilayer perceptron during chip power-on initialization. The miniature multilayer perceptron is configured in a hardware accelerator, which consists of a 16×16-bit multiply-accumulate array, capable of performing 256 multiply-accumulate operations per clock cycle. Within the reserved processing time slots, the hardware accelerator performs forward inference on the fused feature vector. The forward inference process includes matrix multiplication and activation function operations, with the activation function being a modified linear unit function. The time required to complete the forward inference of all layers is less than one hundred microseconds. After forward inference is completed, the output layer outputs a logic vector containing 16 flag bits. The value of each flag bit indicates whether a panel degradation mode is activated. This logic vector is the degradation mode flag bit. The preset panel degradation mode categories include normal state, local pixel aging state, global brightness decay state, Mura non-uniform state, temperature abnormal state, touch noise increase state, etc. For example, when the third bit of the output layer is high, it indicates that the current panel is in the global brightness decay state.

[0041] In one specific embodiment, the weight parameters of the micro multilayer perceptron are obtained through an offline training process: First, a training dataset is constructed, containing multiple fused feature vector samples and their corresponding degradation mode category labels. The fused feature vector samples are obtained by simulating intra-cell touch OLED panel models under different degradation states. These degradation states include at least different degrees of local pixel aging, different degrees of global brightness decay, different degrees of Mura non-uniformity, and different operating temperatures. The corresponding degradation mode category labels use a one-hot encoding format. Second, the loss function is defined as the cross-entropy loss function, and an adaptive moment estimation optimizer is used to train the micro multilayer perceptron. The batch size during training is set to 64, the initial learning rate is set to 0.001, and the learning rate is decayed when the validation set loss no longer decreases. Finally, after training convergence, the obtained weight parameters are programmed into the one-time programmable memory using a programmer.

[0042] The nonlinear compensation model operates in parallel with the micro multilayer sensor and generates a dynamic compensation voltage value according to a preset nonlinear relationship. This dynamic compensation voltage value represents the amount of voltage adjustment applied to the pixel driving circuit required to address pixel degradation caused by organic light-emitting diodes in the touch OLED panel within the compensation unit, and is measured in millivolts.

[0043] The instruction synthesis unit internally maintains a preset gamma correction curve lookup table, which is stored in non-volatile memory. This lookup table contains 16 preset gamma correction curve parameter groups, each corresponding to a gamma correction curve index value. The instruction synthesis unit selects the corresponding gamma correction curve index value from the lookup table based on the combination of high-level flag bits in the degradation mode flag. For example, when the degradation mode flag indicates a global brightness decay state, the selected gamma correction curve index value corresponds to a set of gamma correction parameters that improve the overall display brightness. The selected gamma correction curve index value is then combined and encapsulated with the dynamic compensation voltage value to generate a configuration instruction package.

[0044] In one embodiment of the present invention, in the hybrid reasoning and compensation decision module, the nonlinear compensation model generates a dynamic compensation voltage value. The formula is: in, The average drive current of the second mode signal value, The factory-calibrated reference current value, The temperature rise of the current junction temperature relative to the preset reference temperature. Here, N is the preset temperature acceleration factor, and N is the number of pixels contained within the preset physical block. The current driving current value of the i-th pixel in the second modal signal value. The initial current value of the i-th pixel is the factory current value. This is the scaling factor constant.

[0045] Specifically, the average drive current of the second mode signal value It is the arithmetic mean of the second modal signal values ​​of all pixels composed of organic light-emitting diodes in the touch OLED panel within the display frame corresponding to the current vertical synchronization signal, and the unit is microamps.

[0046] This is the factory-calibrated reference current value stored in non-volatile memory, in microamps. During the chip manufacturing process, measurements are taken under standard testing conditions and stored in non-volatile memory. Standard testing conditions include a typical operating voltage supply, an ambient temperature of 25 degrees Celsius, and a pure white test image displayed on the touch OLED panel within each unit. Under these conditions, the driving current of the pixels composed of organic light-emitting diodes in all touch OLED panels within each unit is measured, and the arithmetic mean is taken as the statistical average. The typical value is 5 microamps.

[0047] The current junction temperature is calculated by subtracting a preset reference temperature value stored in non-volatile memory from the current junction temperature value output by the chip's built-in temperature sensor, expressed in degrees Celsius. The typical preset reference temperature is 25 degrees Celsius, corresponding to the ambient temperature at which the chip was calibrated at the factory. The current junction temperature changes dynamically with the chip's operating status. The value range is typically from 0 degrees Celsius to 60 degrees Celsius; for example, when the current junction temperature of the chip is 55 degrees Celsius, The temperature is 30 degrees Celsius.

[0048] Temperature acceleration factor The unit is per degree Celsius; Based on the pre-calibration of the pixel degradation acceleration characteristics of organic light-emitting diodes in the in-cell touch OLED panel, the relationship between the acceleration factor of 1 degree Celsius increase in temperature and the pixel degradation rate is characterized. The value is based on the activation energy parameter of the organic light-emitting diode material, which is obtained by conducting accelerated aging experiments on the touch OLED panel samples in the unit at different temperatures and fitting the degradation curve. The value ranges from 0.02 per degree Celsius to 0.06 per degree Celsius.

[0049] N represents the number of pixels composed of organic light-emitting diodes in the touch OLED panel within the unit contained in the preset physical block. In a typical embodiment, each physical block corresponds to a 16×16 pixel area on the display panel, therefore N is 256.

[0050] The current driving current value of the pixel composed of organic light-emitting diodes in the i-th unit of the touch OLED panel in the second modal signal value is expressed in microamps. It is acquired and quantized by the multimodal sensing acquisition module within the pixel state sensing window, corresponding to the i-th pixel in the physical block. The value ranges from 1 microamp to 10 microamps.

[0051] The initial factory current value of the pixel composed of organic light-emitting diodes in the i-th unit of the touch OLED panel is expressed in microamps. Measured and stored under standard test conditions during the chip manufacturing process; standard test conditions and The measurement conditions are the same; It reflects the initial electrical characteristics of the pixel at the time of manufacture, and the value ranges from 1 microamp to 10 microamp.

[0052] This is a dimensionless scaling factor constant related to the chip manufacturing process, used to map the formula calculation results to the actual drive voltage compensation range; The value of is determined by the voltage reference accuracy of the chip manufacturing process and the resolution of the digital-to-analog converter, with a typical value of 1.0.

[0053] In one embodiment of the present invention, the frame-level closed-loop update module includes: The bus interface control unit is used to write the configuration instruction packet into the shadow registers of the display driver register group and the touch threshold register group via the internal chip bus within the reserved time slot; The register update control unit is used to load the contents of the shadow register into the working registers of the display driver register group and the touch threshold register group in response to the register update pulse signal generated by the display timing controller when the effective edge of the next vertical synchronization signal after the vertical synchronization signal arrives.

[0054] Specifically, the bus interface control unit is connected to the output buffer of the hybrid inference and compensation decision module via an internal chip bus (e.g., an integrated circuit interface bus), and reads the configuration instruction packet from this output buffer at the start of the reserved time slot. The reserved time slot is provided by the frame synchronization timing control module after the touch sensing window and pixel state sensing window have ended. The start time of the reserved time slot is when the second enable signal becomes inactive, and the end time is when the display blanking interval of the current display frame ends. According to the preset address mapping table, the gamma correction curve index value field in the configuration instruction packet is converted into the write address of the corresponding gamma correction register in the display driver register group, and the dynamic compensation voltage value field is converted into the write address of the corresponding voltage compensation register in the display driver register group and the write address of the corresponding threshold adjustment register in the touch threshold register group.

[0055] In a typical embodiment, the address mapping table contains the following entries: The first entry corresponds to the gamma correction curve index field, which is located at bits 7 to 0 of the configuration instruction packet, a total of 8 bits; the address mapping table maps this field to the address of the gamma correction control register in the display driver register group, which is 0x40 in hexadecimal representation. The second entry corresponds to the dynamic compensation voltage value field, which is located at bits 23 to 8 of the configuration instruction packet, a total of 16 bits; the address mapping table maps this field to two target register addresses simultaneously; the first target address is the address of the voltage compensation register in the display driver register group, which is 0x44 in hexadecimal representation; the second target address is the address of the threshold adjustment register in the touch threshold register group, which is 0x60 in hexadecimal representation.

[0056] Within the reserved time slot, write operations are initiated sequentially to the aforementioned addresses, writing the data of each field to the shadow registers of the display driver register group and the touch threshold register group, respectively. The shadow registers are temporary registers that correspond one-to-one with the working registers; writing to the shadow registers does not immediately affect the display driver parameters and touch detection parameters of the current display frame. The bus write operation is completed before the reserved time slot ends.

[0057] The register update control unit receives a register update pulse signal generated by the display timing controller at the effective edge of the vertical sync signal. When the next effective edge of the vertical sync signal arrives, in response to the register update pulse signal, the contents of the shadow registers of the display drive register group are loaded in parallel into the working registers of the display drive register group, and simultaneously, the contents of the shadow registers of the touch threshold register group are loaded in parallel into the working registers of the touch threshold register group. The parameters in the working registers are directly used to control the display drive voltage and touch detection threshold of the next display frame.

[0058] For example, within the blanking interval of the current display frame, the bus interface control unit writes the gamma correction curve index value 0x03 into the shadow register of the gamma correction register and the dynamic compensation voltage value 0xCF (corresponding to the decimal value 207) into the shadow register of the voltage compensation register. When the current display frame ends and the rising edge of the next vertical sync signal arrives, the register update control unit generates a loading pulse, and the aforementioned shadow register value is loaded into the corresponding working register. The pixel drive voltage of the next display frame then increases by approximately 207 millivolts and uses the gamma correction curve corresponding to index 0x03, thereby completing adaptive compensation for pixel degradation.

[0059] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A frame-synchronized multimodal sensing and intelligent inference system for touch-sensitive OLEDs within a single unit, characterized in that... It includes the following modules: The frame synchronization timing control module is used to respond to the vertical synchronization signal of the display timing controller, and sequentially trigger the touch sensing window and the pixel status sensing window within the display blanking area indicated by the signal, and the sum of the durations of the two windows is less than the preset ratio of the display blanking area; A multimodal sensing acquisition module is used to multiplex the traces of the touch OLED panel within the unit as touch receiving electrodes within the touch sensing window to acquire a first modal analog signal and perform chopping modulation and demodulation on the signal; and to multiplex the traces as detection electrodes within the pixel state sensing window to acquire a second modal analog signal. The on-chip feature extraction and fusion module is used to acquire the current junction temperature of the chip and perform temperature compensation on the first modal signal value obtained by quantizing the first modal analog signal according to a preset temperature coefficient. It also divides the first modal signal value and the second modal signal value obtained by quantizing the second modal analog signal into preset physical blocks, calculates the statistical deviation feature of the temperature-compensated first modal signal value relative to the historical baseline value in the non-volatile memory and the degradation deviation feature of the second modal signal value relative to the factory initial value in the non-volatile memory for each physical block, and fuses the statistical deviation feature and degradation deviation feature based on the grayscale value of the display frame corresponding to the vertical synchronization signal to generate a fused feature vector. The hybrid reasoning and compensation decision module incorporates a micro multilayer perceptron and a nonlinear compensation model. The micro multilayer perceptron obtains a degradation mode flag based on the fused feature vector. The nonlinear compensation model generates a dynamic compensation voltage value based on the average drive current of the second mode signal value, the temperature rise of the current junction temperature relative to a preset reference temperature, the degradation deviation characteristics, and the preset temperature acceleration factor. A configuration instruction package is generated based on the degradation mode flag and the dynamic compensation voltage value. The frame-level closed-loop update module is used to write the configuration instruction package into the display driver register and the touch threshold register within a reserved time slot other than the sum of the durations of the touch sensing window and the pixel state sensing window.

2. The multimodal sensing and intelligent inference system for frame synchronization of an in-unit touch OLED according to claim 1, characterized in that... The frame synchronization timing control module includes a duration constraint judgment logic. This duration constraint judgment logic accumulates the configuration duration of the touch sensing window and the configuration duration of the pixel state sensing window, and compares the accumulated value with a preset ratio threshold of the display blanking area. When the accumulated value exceeds the preset ratio threshold, the configuration duration of the touch sensing window and the pixel state sensing window is truncated.

3. The multimodal sensing and intelligent inference system for frame synchronization of an in-unit touch OLED according to claim 1, characterized in that... The multimodal sensing acquisition module includes: An analog switch matrix has its input terminals connected to the traces of the touch OLED panel within the unit, and its output terminals connected to the touch sensing circuit within the touch sensing window to acquire the original first-mode analog signal, and connected to the pixel state detection circuit within the pixel state sensing window to acquire the second-mode analog signal; The touch sensing circuit includes a chopping modulation unit for modulating the original first-mode analog signal to a preset carrier frequency; and a synchronous demodulation unit for demodulating the modulated original first-mode analog signal using a reference signal that is in phase and frequency with the modulation end, and outputting the demodulated signal as the first-mode analog signal. The pixel state detection circuit includes a test signal generation unit for applying a test electrical signal to the trace within the pixel state sensing window to drive the pixel to conduct; and a transimpedance amplification unit for converting the current generated by the pixel conduction into a voltage signal to complete the acquisition of the second mode analog signal.

4. The multimodal sensing and intelligent inference system for frame synchronization of an in-unit touch OLED according to claim 1, characterized in that... In the on-chip feature extraction and fusion module, the current junction temperature of the chip is acquired, and temperature compensation is performed on the first modal signal value obtained by quantization of the first modal analog signal according to a preset temperature coefficient, including: Based on the current junction temperature of the chip, the preset reference temperature, the preset temperature coefficient, and the preset linear relationship between the first modal signal value and the first modal signal value, the temperature-compensated first modal signal value is calculated after the touch sensing window of the display frame corresponding to the vertical synchronization signal ends and before the statistical deviation feature is calculated; wherein the preset temperature coefficient is a coefficient value pre-calibrated for the capacitance temperature drift characteristics of the touch OLED panel traces in the unit.

5. The multimodal sensing and intelligent inference system for frame synchronization of an in-unit touch OLED according to claim 1, characterized in that... The on-chip feature extraction and fusion module further includes: The statistical deviation feature generation unit is used to perform a difference calculation between the standard deviation of the temperature-compensated first modal signal value in the display frame corresponding to the vertical synchronization signal and the standard deviation of the historical baseline collected under no-touch conditions during the chip initialization phase, to obtain the statistical deviation feature. The degradation deviation feature generation unit is used to calculate the root mean square value of the relative deviation of the second mode signal value with respect to the initial value collected under standard test conditions during the chip manufacturing stage, and to obtain the degradation deviation feature; The feature fusion unit is used to determine the weighting coefficient based on the grayscale value of each physical block corresponding to the display frame corresponding to the vertical synchronization signal, and to fuse the difference of the standard deviation and the root mean square value of the relative deviation with the weighting coefficient to obtain the fused feature vector.

6. The multimodal sensing and intelligent inference system for frame synchronization of an in-unit touch OLED according to claim 1, characterized in that... The hybrid reasoning and compensation decision-making module includes: The miniature multilayer perceptron consists of a fully connected input layer, hidden layer, and output layer. Its weight parameters are pre-programmed in a one-time programmable memory. The miniature multilayer perceptron is configured in a hardware accelerator and performs forward inference on the fused feature vector within the reserved processing time slot of the display frame corresponding to the vertical synchronization signal, and outputs a degradation mode flag. A nonlinear compensation model generates a dynamic compensation voltage value based on the average drive current of the second mode signal value, the temperature rise of the current junction temperature relative to a preset reference temperature, the degradation deviation characteristics, and the preset temperature acceleration factor. The instruction synthesis unit selects the corresponding gamma correction curve index value from the preset gamma correction curve lookup table according to the degradation mode flag, and combines the gamma correction curve index value with the dynamic compensation voltage value to form a configuration instruction package.

7. The multimodal sensing and intelligent inference system for frame synchronization of an in-unit touch OLED according to claim 6, characterized in that... In the hybrid reasoning and compensation decision-making module, the nonlinear compensation model generates dynamic compensation voltage values. The formula is: in, The average drive current of the second mode signal value, The factory-calibrated reference current value, The temperature rise of the current junction temperature relative to the preset reference temperature. Here, N is the preset temperature acceleration factor, and N is the number of pixels contained within the preset physical block. The current driving current value of the i-th pixel in the second modal signal value. The initial current value of the i-th pixel is the factory current value. This is the scaling factor constant.

8. The multimodal sensing and intelligent inference system for frame synchronization of an in-unit touch OLED according to claim 1, characterized in that... The frame-level closed-loop update module includes: The bus interface control unit is used to write the configuration instruction packet into the shadow registers of the display driver register group and the touch threshold register group via the internal chip bus within the reserved time slot; The register update control unit is used to load the contents of the shadow register into the working registers of the display driver register group and the touch threshold register group in response to the register update pulse signal generated by the display timing controller when the effective edge of the next vertical synchronization signal after the vertical synchronization signal arrives.

Images